Gastric cancer (GC) is one of the most virulent cancers in the world. Each year over 930,000 of people are diagnosed with GC and 700,000 die (2nd leading cause of death) worldwide. Patients with localized gastric cancer are most often treated surgery alone (in Asia) or surgery plus adjunct therapy (North America). This decision for treatment is empiric. There is no method to select a specific therapy or to avoid a specific one. Therefore, nearly all patients with AJCC/UICC stage II or III cancer are offered a similar treatment but their outcomes are unpredictable. We have considerable preliminary data that suggests a PCR-based practical prognostic signature (that allows us to compute a continuous "risk" score) for all pathologic stages is possible. Our main objective is to develop a PCR-based biomarker prognostic test for an "early" (= 12 months) recurrence-free survival (RFS) event for localized gastric cancer patients who undergo surgery with or without adjunct therapy. If the emerging prognostic signature has a high rate of accuracy (= 80%), alternate strategies can be developed to optimize treatment. Currently, we cannot optimize therapy and end up giving the same therapy to all patients with similar stage. Our overall goal is to validate previously identified prognostic gene signature in larger independent Asian GC patient cohort and Western GC patient cohorts, and improve the predictability of signatures by identifying more biomarkers to construct better prognostic signature and risk score. PUBLIC HEALTH RELEVANCE: The overall goal of this proposal is to develop novel prognostic risk score that can predict the likelihood of early recurrence in gastric cancer patients. Our approach will pave the way for the optimization of patient treatments according to relative risk of patients. Currently, we cannot optimize therapy and end up giving the same therapy to all patients with similar stage.